Re: Services I'd like from auric
On Fri, Jan 09, 2004 at 02:14:53PM -0500, Branden Robinson wrote:
> On Thu, Jan 01, 2004 at 09:44:50PM +0000, Colin Watson wrote:
> > On Thu, Jan 01, 2004 at 04:36:56PM -0500, Branden Robinson wrote:
> > > On Wed, Dec 31, 2003 at 03:30:03AM +0000, Colin Watson wrote:
> > > > On Tue, Dec 30, 2003 at 03:57:09PM -0500, Branden Robinson wrote:
> > > > > Do we still have cron jobs running stuff out of people's home
> > > > > directories? If so, shouldn't they be changed to execute out of the
> > > > > publicly accessible directories, or, better yet, the requisite stuff
> > > > > actually checked into CVS?
> > > >
> > > > It's my understanding that the URL above is mapped to a directory in
> > > > /org/ftp.debian.org on auric.
> > >
> > > Perhaps I'm dense, but I don't see how that answers my questions.
> > Well, it indicates that the testing scripts no longer run out of ~ajt,
> > but from /org/ftp.debian.org, which I thought was what you were talking
> > about ...?
> I wasn't talking about just the testing scripts. I was talking about
> all administrative cron jobs of any nature running out of people's home
Oh, I see. Er, probably "yes", then, but I'm not sure that my saying
that helps you any. :)
There's often a rather blurred boundary between "stuff I'm experimenting
with" and "part of a service", and most code that's part of a debian.org
service crosses that boundary at some point in its life. While it's in
the experimental phase, it's entirely appropriate for it to live in
somebody's home directory, and sometimes people just forget to move it,
or leave it be on a half-thought basis of "well, it works the way it is
and now I have more urgent things to look at". Extensions to the bug
system indexes are a particularly good example of this: they've always
been experimental, nobody's ever really settled on something
well-designed and solid enough to check into CVS for general use by
users of the debbugs package, and yet because they're useful they have a
habit of creeping into production use on bugs.debian.org and making life
easier for the users of that service. It's not the best situation in the
world, but it's better than the experimentation not happening at all.
I guess what I'm saying is that yes, I want administrative cron jobs to
live in shared directories, be writable by the relevant group, be
checked into CVS, and so on; but I don't think it's always practical to
set a hard-and-fast rule to that effect for evolving services. We just
have to handle it on a case-by-case basis.
Colin Watson [email@example.com]